The first question to be struck by the campus network NATSHELLL_AUTH_BILLING system is not which manufacturer or model to buy, but the mode of operation:Is it a proxy or self-build model?If this decision is wrong, the network structure behind it, the device selection, the docking development, and the operator coordination all go wrong. But in the actual project, many schools are passively decided on this node.——The operator came first, and left on which calibre, without any real model comparison.
The core difference between the proxy model and the self-building model is not at the technical level, but rather at the very least.Who pays, who charges, who runs, who takes responsibility?These four questions. It's much more useful to get these questions straight, than to study a bunch of technical parameters. Many information centre teachers talk to me and say that the most confusing thing they have in their selection phase is this: the market looks like a similar scheme, and the parameters in the tender are very different, but they really run away from it.
The proxy model is essentially to allow operators to export the Internet. Schools build basic networks (wireless APs, cable access, core switchboards), and operators have broadband accounts connected to the campus exports through dial-up gateways. Students enter a campus network account on Portal’s page, but the dial-up gateway converts it into an operator wide account at backstage, initiating PPPE dialling to the operator Bras.——If you say so, the information centre is small and students are going to ask about it every day."Why is the net slow?""Why can't you get on the Internet with all that money?"And they can't handle the billing business.
The model is completely different. Schools deploy the entire Authentication & Billing system, defining their own set of meals, pricing themselves, charging for their own fees. Students access it through Portal, 802.1X or PPOE, and then use the Internet bandwidth that they pull themselves on.The school has full autonomy.——Prices are priced as they are, and the colleges that you want to give preferential treatment are not subject to operators. But the cost is also obvious: schools deal with their own petty matters of fees, refunds, complaints, reconciliations. Information centres have become half-operators from a purely technical sector.
The most common pit I've seen in the actual project is to confuse the boundaries of these two models. Some schools want to make a sub-division, but they say to operators at the same time:"Our own Portal page shows the school bulletins and curriculum."And then the operator said,"Our Bras doesn't support this complicated Portal delivery."There are schools that build themselves, but forget to think about a basic question: Is the school’s own Internet export bandwidth enough? A 20,000 school, if all of them go in one way or another, could reach 8,000 students at night and send it online. It is impossible to do so by adding a gigabyte of operator broadband to the school’s educational network alone. These issues are discussed too late when you choose to buy hardware.
One of the most central technological dependence of the dialysis model isIntercompatibility of dial-up gateways and operator BrasTheoretically, the standard RADIAUS protocol is not a problem, but the software version and model of the company operator Bras varies. For example, there are slight differences in the treatment of the RADIAS properties field between bras and H3C Bras——Some attributes are not supported by H3C, or the same properties are disassembly derived from two platforms. More problematic is that some local operators' BRAS are old versions of deployments many years ago, and even RADIUS’s specific properties are not supported.
The core difficulty of the self-building model is thatIntegration of costing logic and campus operations systemSchools are not just operators, and students are not just Internet users. The identity of a student in the campus network Authentication & Billing must be consistent with his status in the teaching system, a cartoon system, library systems."The teaching service graduated and retired, but Authentication & Billing system is still charging him for the Internet."This is a U-Lone. Don't think it's a small probability event.——It is not realistic to check write-offs by manual, with 20,000 students per year enrolled and graduating. Automatic synchronization between systems must be done through LDAP or API.
And there's another variable that is often overlooked.Multi-operator coexistenceMany schools now have a tendency not to monopolize but to combine the three sets of connectivity, mobile and telecommunications. Students choose which one they want. In this case, the dial-up gateway must be able to handle BRAS from three operators simultaneously and can automatically select the corresponding operator line based on the user-bound package. This function is much more difficult to achieve than single operator allocation.——The BRAS configuration specifications, dial-up parameters and AAA returns properties of different operators vary, and the dial-up gateway requires a well developed multi-operator compatibility capability.
There is a compromise formula in practice:The school built itself the Authentication & Billing platform for local certification and user management, with Internet export operators providing proxy.The logic of this program is that the school NATHELL_AUTH_BILLING system has been certified at first level (account number, suite, access), and if the user binds the operator's account number, then it makes secondary certification and sub-dealing to the operator BRAS through a dial-up gateway. The advantage is that schools retain control over user data, authentication strategies and Portal pages, but dump the baggage on export bandwidth to operators. The cost is two layers of certification process, increasing complexity of the system, and more complex interface between the depotrator gateway and the certification platform than simple matching.
Finally, a point that many people do not appreciate but have a great impact in practice:Operators ' cooperationThe mode of sub-opt is largely dependent on the willingness and speed of operators to cooperate. In some provinces, operators are very reluctant to do business with campus sub-placing.——Because the unit price of a student package is low and ARPU is not high, operators find input less productive. In this case, schools are likely to be held up by operators, with project cycles more than double that expected. If there is no school plan B on this issue, they can only do it. So when you choose a model, you must assess the actual ability and willingness of the operator to cooperate, not just whether the programme is good or not. This is not a purely technical problem, but often more difficult than technology.
In summary, there is no absolute right or wrong to choose between the models of proxy and self-building, except for the difference between fit and unsuitable. The criteria are three:Are schools willing and able to operate autonomously? Who will take the lead on the issue of export bandwidth? Can the actual cooperation of operators support the project's normal progress?It would be more useful to write down the answers to those three questions than to read 10 programme books.